Deepfakes [UPSC Notes]

Deepfakes is a term frequently in the news nowadays. It signifies the endless possibilities of technology and also presents the dark side of unhinged technology. What are deepfakes? How harmful can they be? Read on to know more about deepfakes. This is part of the UPSC syllabus science & technology segment.

What are Deepfakes?

Deepfakes are digital media – video, audio and images edited and manipulated using Artificial Intelligence. This synthetic media content is referred to as ‘deepfakes’. It is basically hyper-realistic falsification.

• The name comes from a combination of two terms – ‘deep learning’ and ‘fake’. Deep learning is a type of machine learning based on Artificial Neural Networks (ANNs).
•  Artificial Intelligence (AI) generated synthetic media or deep fakes offer a number of benefits in certain areas like film production, education, artistic expression and criminal forensics.
• Cybercriminals are using deepfake technology to carry out phishing attacks, financial fraud, identity theft, information manipulation and also political unrest. 

The name ‘deepfakes’ originated in 2017 when a Reddit user by the name ‘deepfakes’, posted explicit videos of celebrities. 

How Deepfakes Work?

A Deepfake video seems like original content having the person doing some kind of action or speaking on a topic. And while creating such fake videos, multiple images of the targeted person from different angles are superimposed on the original face.

• The developers of deepfake software typically use GANs (Generative Adversarial Networks).
• GAN consists of two neural networks – the generator that creates fake media and the discriminator that determines whether the media is fake or real.
• Every time the discriminator correctly recognises the content as being fake, it gives the generator important insights into how to make the next deep fakes better.
• Deepfake generation techniques usually require many images and videos of the targets. High-profile persons like celebrities, and political leaders become easy targets because of the large data sets available on the internet. Thus, deep fakes can seriously harm the reputation of any high-profile person.

Deepfakes Concerns

• One of the serious threats from deepfake technology is synthetic adult material/deep-fake porn.
• Anyone can create and publish a video so realistic that it will have the potential to harm people, organisations and societies.
• Sometimes an undesirable truth may be treated as a deepfake.
• Deepfakes can be used by non-state actors to stir anti-state sentiments among people. 
• Deepfakes can be used for sabotaging democratic processes like elections. 

How to deal with Deepfakes?

Detection – Building technologies to identify deepfakes

• Big tech firms like Google, Meta and Microsoft have condemned deepfake technology and are creating tools to recognise them.
• Microsoft is creating a new anti-deep fake technology to fight misinformation called Microsoft Video Authenticator.

Transparency

• To instil trust, social media companies created “verified accounts”.
• One proposition to combat deep fakes points to the usage of NFT (Non-Fungible Tokens) as a possible solution. 

Regulations

• In 2019, there were about a dozen Federal and State Level Bills in the US to regulate deep fakes. These laws range from criminalising the use of a woman’s likeness in a pornographic film without her consent (Virginia Law), to the appropriate use of a deceased person’s data (New York Law), and dealing with cheap fakes (low-tech digital frauds not requiring AI).
• A proper regulatory framework can facilitate disincentivizing the creation and distribution of malicious deep fakes.

Education

• Media literacy for consumers is the most effective tool to combat disinformation and deep fakes.
• The greater public awareness of the technology and its uses, the more they will be able to think critically about the media they consume and apply caution where needed.

Deepfakes:- Download PDF Here